Team Pro

Team Pro is designed for an eight-person team with one or two active coding agents alongside business users. That is the intended mix, not an unlimited number of simultaneous generations. Workloads and context limits are agreed during setup.

The workstation is for demanding coding agents, shared inference, and CUDA workflows: repository work that can stay on your network, document search, and the everyday drafting and operations tasks the rest of the team needs. Dedicated GPU memory is a different resource from the unified memory in the Mac boxes, with different capacity and performance characteristics.

We select the model and quantization for these GPUs and qualify new releases before adding them. The current evaluation set includes Qwen3.8-27B, Qwen3.6-35B-A3B, and Qwen3-Coder-Next. Larger models and longer context need more memory and may reduce concurrency.

Team Pro is a Linux workstation built around two NVIDIA RTX PRO 6000 Blackwell Max-Q GPUs, 96 GB of GPU memory each, a Threadripper 9960X, 256 GB of ECC system memory, and 8 TB of NVMe storage.

The NVIDIA system uses a CUDA inference stack selected and tested for its models. Expansion depends on the motherboard, chassis, power, and cooling. We document the supported upgrades with your configuration.